[IGNITE-8057] Execute WAL fsync preventively before checkpoint begin

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.5
• Assignee: Unassigned
• Reporter: Ivan Rakov
• Created: 27/Mar/18 11:51
• Updated: 27/Mar/18 12:17

Description

After fix of IGNITE-7754, we execute explicit WAL fsync on checkpoint begin:

if (hasPages) {
    assert cpPtr != null;

    tracker.onWalCpRecordFsyncStart();

    // Sync log outside the checkpoint write lock.
    cctx.wal().flush(cpPtr, true);

    tracker.onWalCpRecordFsyncEnd();

It's executed outside of checkpoint write lock. However, it still can decrease overall throughput by suspending writing of dirty pages by checkpoint threads.
We can decrease time of this fsync by executing it preventively, before acquiring checkpoint write lock.
We should prioritize this ticket if value of walCpRecordFsyncDuration metric in "Checkpoint started" message will be too big.
Note: it's possible to give a fsync hint to WAL manager in single-writer mode.

Issue Links

• is caused by: IGNITE-7754 WAL in LOG_ONLY mode doesn't execute fsync on checkpoint begin